Examples with StorageFormat com.facebook.presto.hive.metastore.StorageFormat used on opensource projects

Example 11 with StorageFormat

use of com.facebook.presto.hive.metastore.StorageFormat in project presto by prestodb.

the class RcFileFileWriterFactory method createFileWriter.

@Override
public Optional<HiveFileWriter> createFileWriter(Path path, List<String> inputColumnNames, StorageFormat storageFormat, Properties schema, JobConf configuration, ConnectorSession session) {
    if (!HiveSessionProperties.isRcfileOptimizedWriterEnabled(session)) {
        return Optional.empty();
    }
    if (!RCFileOutputFormat.class.getName().equals(storageFormat.getOutputFormat())) {
        return Optional.empty();
    }
    RcFileEncoding rcFileEncoding;
    if (LazyBinaryColumnarSerDe.class.getName().equals(storageFormat.getSerDe())) {
        rcFileEncoding = new BinaryRcFileEncoding();
    } else if (ColumnarSerDe.class.getName().equals(storageFormat.getSerDe())) {
        rcFileEncoding = createTextVectorEncoding(schema, hiveStorageTimeZone);
    } else {
        return Optional.empty();
    }
    Optional<String> codecName = Optional.ofNullable(configuration.get(FileOutputFormat.COMPRESS_CODEC));
    // existing tables and partitions may have columns in a different order than the writer is providing, so build
    // and index to rearrange columns in the proper order
    List<String> fileColumnNames = Splitter.on(',').trimResults().omitEmptyStrings().splitToList(schema.getProperty(META_TABLE_COLUMNS, ""));
    List<Type> fileColumnTypes = toHiveTypes(schema.getProperty(META_TABLE_COLUMN_TYPES, "")).stream().map(hiveType -> hiveType.getType(typeManager)).collect(toList());
    int[] fileInputColumnIndexes = fileColumnNames.stream().mapToInt(inputColumnNames::indexOf).toArray();
    try {
        FileSystem fileSystem = hdfsEnvironment.getFileSystem(session.getUser(), path, configuration);
        OutputStream outputStream = fileSystem.create(path);
        Optional<Supplier<RcFileDataSource>> validationInputFactory = Optional.empty();
        if (HiveSessionProperties.isRcfileOptimizedWriterValidate(session)) {
            validationInputFactory = Optional.of(() -> {
                try {
                    return new HdfsRcFileDataSource(path.toString(), fileSystem.open(path), fileSystem.getFileStatus(path).getLen());
                } catch (IOException e) {
                    throw Throwables.propagate(e);
                }
            });
        }
        return Optional.of(new RcFileFileWriter(outputStream, rcFileEncoding, fileColumnTypes, codecName, fileInputColumnIndexes, ImmutableMap.<String, String>builder().put(HiveMetadata.PRESTO_VERSION_NAME, nodeVersion.toString()).put(HiveMetadata.PRESTO_QUERY_ID_NAME, session.getQueryId()).build(), validationInputFactory));
    } catch (Exception e) {
        throw Throwables.propagate(e);
    }
}

Example 12 with StorageFormat

use of com.facebook.presto.hive.metastore.StorageFormat in project presto by prestodb.

the class HiveMetadata method computePartitionUpdatesForMissingBuckets.

private List<PartitionUpdate> computePartitionUpdatesForMissingBuckets(ConnectorSession session, HiveWritableTableHandle handle, Table table, List<PartitionUpdate> partitionUpdates) {
    // avoid creation of PartitionUpdate with empty list of files
    if (!shouldCreateFilesForMissingBuckets(table, session)) {
        return ImmutableList.of();
    }
    HiveStorageFormat storageFormat = table.getPartitionColumns().isEmpty() ? handle.getTableStorageFormat() : handle.getPartitionStorageFormat();
    // empty un-partitioned bucketed table
    if (table.getPartitionColumns().isEmpty() && partitionUpdates.isEmpty()) {
        int bucketCount = handle.getBucketProperty().get().getBucketCount();
        LocationHandle locationHandle = handle.getLocationHandle();
        List<String> fileNamesForMissingBuckets = computeFileNamesForMissingBuckets(session, storageFormat, handle.getCompressionCodec(), bucketCount, ImmutableSet.of());
        return ImmutableList.of(new PartitionUpdate("", (handle instanceof HiveInsertTableHandle) ? APPEND : NEW, locationHandle.getWritePath(), locationHandle.getTargetPath(), fileNamesForMissingBuckets.stream().map(fileName -> new FileWriteInfo(fileName, fileName, Optional.of(0L))).collect(toImmutableList()), 0, 0, 0, isFileRenamingEnabled(session)));
    }
    ImmutableList.Builder<PartitionUpdate> partitionUpdatesForMissingBucketsBuilder = ImmutableList.builder();
    for (PartitionUpdate partitionUpdate : partitionUpdates) {
        int bucketCount = handle.getBucketProperty().get().getBucketCount();
        List<String> fileNamesForMissingBuckets = computeFileNamesForMissingBuckets(session, storageFormat, handle.getCompressionCodec(), bucketCount, ImmutableSet.copyOf(getTargetFileNames(partitionUpdate.getFileWriteInfos())));
        partitionUpdatesForMissingBucketsBuilder.add(new PartitionUpdate(partitionUpdate.getName(), partitionUpdate.getUpdateMode(), partitionUpdate.getWritePath(), partitionUpdate.getTargetPath(), fileNamesForMissingBuckets.stream().map(fileName -> new FileWriteInfo(fileName, fileName, Optional.of(0L))).collect(toImmutableList()), 0, 0, 0, isFileRenamingEnabled(session)));
    }
    return partitionUpdatesForMissingBucketsBuilder.build();
}

Example 13 with StorageFormat

use of com.facebook.presto.hive.metastore.StorageFormat in project presto by prestodb.

the class ParquetFileWriterFactory method createFileWriter.

@Override
public Optional<HiveFileWriter> createFileWriter(Path path, List<String> inputColumnNames, StorageFormat storageFormat, Properties schema, JobConf conf, ConnectorSession session, Optional<EncryptionInformation> encryptionInformation) {
    if (!isParquetOptimizedWriterEnabled(session)) {
        return Optional.empty();
    }
    if (!MapredParquetOutputFormat.class.getName().equals(storageFormat.getOutputFormat())) {
        return Optional.empty();
    }
    ParquetWriterOptions parquetWriterOptions = ParquetWriterOptions.builder().setMaxPageSize(getParquetWriterPageSize(session)).setMaxBlockSize(getParquetWriterBlockSize(session)).build();
    CompressionCodecName compressionCodecName = getCompression(conf);
    List<String> fileColumnNames = Splitter.on(',').trimResults().omitEmptyStrings().splitToList(schema.getProperty(META_TABLE_COLUMNS, ""));
    List<Type> fileColumnTypes = toHiveTypes(schema.getProperty(META_TABLE_COLUMN_TYPES, "")).stream().map(hiveType -> hiveType.getType(typeManager)).collect(toList());
    int[] fileInputColumnIndexes = fileColumnNames.stream().mapToInt(inputColumnNames::indexOf).toArray();
    try {
        FileSystem fileSystem = hdfsEnvironment.getFileSystem(session.getUser(), path, conf);
        Callable<Void> rollbackAction = () -> {
            fileSystem.delete(path, false);
            return null;
        };
        return Optional.of(new ParquetFileWriter(fileSystem.create(path), rollbackAction, fileColumnNames, fileColumnTypes, parquetWriterOptions, fileInputColumnIndexes, compressionCodecName));
    } catch (IOException e) {
        throw new PrestoException(HIVE_WRITER_OPEN_ERROR, "Error creating Parquet file", e);
    }
}

Example 14 with StorageFormat

use of com.facebook.presto.hive.metastore.StorageFormat in project presto by prestodb.

the class TestHivePageSourceProvider method testUseRecordReaderWithInputFormatAnnotationAndCustomSplit.

@Test
public void testUseRecordReaderWithInputFormatAnnotationAndCustomSplit() {
    StorageFormat storageFormat = StorageFormat.create(ParquetHiveSerDe.class.getName(), HoodieParquetInputFormat.class.getName(), "");
    Storage storage = new Storage(storageFormat, "test", Optional.empty(), true, ImmutableMap.of(), ImmutableMap.of());
    Map<String, String> customSplitInfo = ImmutableMap.of(CUSTOM_FILE_SPLIT_CLASS_KEY, HoodieRealtimeFileSplit.class.getName(), HUDI_BASEPATH_KEY, "/test/file.parquet", HUDI_DELTA_FILEPATHS_KEY, "/test/.file_100.log", HUDI_MAX_COMMIT_TIME_KEY, "100");
    HiveRecordCursorProvider recordCursorProvider = new MockHiveRecordCursorProvider();
    HiveBatchPageSourceFactory hiveBatchPageSourceFactory = new MockHiveBatchPageSourceFactory();
    Optional<ConnectorPageSource> pageSource = HivePageSourceProvider.createHivePageSource(ImmutableSet.of(recordCursorProvider), ImmutableSet.of(hiveBatchPageSourceFactory), new Configuration(), new TestingConnectorSession(new HiveSessionProperties(new HiveClientConfig().setUseRecordPageSourceForCustomSplit(true), new OrcFileWriterConfig(), new ParquetFileWriterConfig(), new CacheConfig()).getSessionProperties()), new Path("/test/"), OptionalInt.empty(), 0, 100, 200, Instant.now().toEpochMilli(), storage, TupleDomain.none(), ImmutableList.of(), ImmutableMap.of(), ImmutableList.of(), DateTimeZone.UTC, new TestingTypeManager(), new SchemaTableName("test", "test"), ImmutableList.of(), ImmutableList.of(), ImmutableMap.of(), 0, TableToPartitionMapping.empty(), Optional.empty(), false, null, null, false, null, Optional.empty(), customSplitInfo);
    assertTrue(pageSource.isPresent());
    assertTrue(pageSource.get() instanceof RecordPageSource);
}

Example 15 with StorageFormat

use of com.facebook.presto.hive.metastore.StorageFormat in project presto by prestodb.

the class HiveSplitManager method getPartitionMetadata.

private Iterable<HivePartitionMetadata> getPartitionMetadata(SemiTransactionalHiveMetastore metastore, Table table, SchemaTableName tableName, List<HivePartition> hivePartitions, Optional<HiveBucketHandle> hiveBucketHandle, ConnectorSession session, WarningCollector warningCollector, Optional<Set<HiveColumnHandle>> requestedColumns, Map<String, HiveColumnHandle> predicateColumns, Optional<Map<Subfield, Domain>> domains) {
    if (hivePartitions.isEmpty()) {
        return ImmutableList.of();
    }
    Optional<Set<HiveColumnHandle>> allRequestedColumns = mergeRequestedAndPredicateColumns(requestedColumns, ImmutableSet.copyOf(predicateColumns.values()));
    if (hivePartitions.size() == 1) {
        HivePartition firstPartition = getOnlyElement(hivePartitions);
        if (firstPartition.getPartitionId().equals(UNPARTITIONED_ID)) {
            return ImmutableList.of(new HivePartitionMetadata(firstPartition, Optional.empty(), TableToPartitionMapping.empty(), encryptionInformationProvider.getReadEncryptionInformation(session, table, allRequestedColumns), ImmutableSet.of()));
        }
    }
    StorageFormat storageFormat = table.getStorage().getStorageFormat();
    Optional<HiveStorageFormat> hiveStorageFormat = getHiveStorageFormat(storageFormat);
    Optional<HiveStorageFormat> resolvedHiveStorageFormat;
    if (isUseParquetColumnNames(session)) {
        // Use Hive Storage Format as Parquet if table is of HUDI format
        resolvedHiveStorageFormat = (!hiveStorageFormat.isPresent() && isHudiFormat(storageFormat)) ? Optional.of(PARQUET) : hiveStorageFormat;
    } else {
        resolvedHiveStorageFormat = hiveStorageFormat;
    }
    Iterable<List<HivePartition>> partitionNameBatches = partitionExponentially(hivePartitions, minPartitionBatchSize, maxPartitionBatchSize);
    Iterable<List<HivePartitionMetadata>> partitionBatches = transform(partitionNameBatches, partitionBatch -> {
        Map<String, PartitionSplitInfo> partitionSplitInfo = getPartitionSplitInfo(session, metastore, tableName, partitionBatch, predicateColumns, domains);
        if (partitionBatch.size() != partitionSplitInfo.size()) {
            throw new PrestoException(GENERIC_INTERNAL_ERROR, format("Expected %s partitions but found %s", partitionBatch.size(), partitionSplitInfo.size()));
        }
        Map<String, Partition> partitions = partitionSplitInfo.entrySet().stream().collect(toImmutableMap(Entry::getKey, entry -> entry.getValue().getPartition()));
        Optional<Map<String, EncryptionInformation>> encryptionInformationForPartitions = encryptionInformationProvider.getReadEncryptionInformation(session, table, allRequestedColumns, partitions);
        ImmutableList.Builder<HivePartitionMetadata> results = ImmutableList.builder();
        Map<String, Set<String>> partitionsNotReadable = new HashMap<>();
        int unreadablePartitionsSkipped = 0;
        for (HivePartition hivePartition : partitionBatch) {
            Partition partition = partitions.get(hivePartition.getPartitionId());
            if (partitionSplitInfo.get(hivePartition.getPartitionId()).isPruned()) {
                continue;
            }
            if (partition == null) {
                throw new PrestoException(GENERIC_INTERNAL_ERROR, "Partition not loaded: " + hivePartition);
            }
            String partitionName = makePartName(table.getPartitionColumns(), partition.getValues());
            Optional<EncryptionInformation> encryptionInformation = encryptionInformationForPartitions.map(metadata -> metadata.get(hivePartition.getPartitionId()));
            if (!isOfflineDataDebugModeEnabled(session)) {
                // verify partition is online
                verifyOnline(tableName, Optional.of(partitionName), getProtectMode(partition), partition.getParameters());
                // verify partition is not marked as non-readable
                String reason = partition.getParameters().get(OBJECT_NOT_READABLE);
                if (!isNullOrEmpty(reason)) {
                    if (!shouldIgnoreUnreadablePartition(session) || !partition.isEligibleToIgnore()) {
                        throw new HiveNotReadableException(tableName, Optional.of(partitionName), reason);
                    }
                    unreadablePartitionsSkipped++;
                    if (partitionsNotReadable.size() <= 3) {
                        partitionsNotReadable.putIfAbsent(reason, new HashSet<>(ImmutableSet.of(partitionName)));
                        if (partitionsNotReadable.get(reason).size() <= 3) {
                            partitionsNotReadable.get(reason).add(partitionName);
                        }
                    }
                    continue;
                }
            }
            // Verify that the partition schema matches the table schema.
            // Either adding or dropping columns from the end of the table
            // without modifying existing partitions is allowed, but every
            // column that exists in both the table and partition must have
            // the same type.
            List<Column> tableColumns = table.getDataColumns();
            List<Column> partitionColumns = partition.getColumns();
            if ((tableColumns == null) || (partitionColumns == null)) {
                throw new PrestoException(HIVE_INVALID_METADATA, format("Table '%s' or partition '%s' has null columns", tableName, partitionName));
            }
            TableToPartitionMapping tableToPartitionMapping = getTableToPartitionMapping(session, resolvedHiveStorageFormat, tableName, partitionName, tableColumns, partitionColumns);
            if (hiveBucketHandle.isPresent() && !hiveBucketHandle.get().isVirtuallyBucketed()) {
                Optional<HiveBucketProperty> partitionBucketProperty = partition.getStorage().getBucketProperty();
                if (!partitionBucketProperty.isPresent()) {
                    throw new PrestoException(HIVE_PARTITION_SCHEMA_MISMATCH, format("Hive table (%s) is bucketed but partition (%s) is not bucketed", hivePartition.getTableName(), hivePartition.getPartitionId()));
                }
                int tableBucketCount = hiveBucketHandle.get().getTableBucketCount();
                int partitionBucketCount = partitionBucketProperty.get().getBucketCount();
                List<String> tableBucketColumns = hiveBucketHandle.get().getColumns().stream().map(HiveColumnHandle::getName).collect(toImmutableList());
                List<String> partitionBucketColumns = partitionBucketProperty.get().getBucketedBy();
                if (!tableBucketColumns.equals(partitionBucketColumns) || !isBucketCountCompatible(tableBucketCount, partitionBucketCount)) {
                    throw new PrestoException(HIVE_PARTITION_SCHEMA_MISMATCH, format("Hive table (%s) bucketing (columns=%s, buckets=%s) is not compatible with partition (%s) bucketing (columns=%s, buckets=%s)", hivePartition.getTableName(), tableBucketColumns, tableBucketCount, hivePartition.getPartitionId(), partitionBucketColumns, partitionBucketCount));
                }
            }
            results.add(new HivePartitionMetadata(hivePartition, Optional.of(partition), tableToPartitionMapping, encryptionInformation, partitionSplitInfo.get(hivePartition.getPartitionId()).getRedundantColumnDomains()));
        }
        if (unreadablePartitionsSkipped > 0) {
            StringBuilder warningMessage = new StringBuilder(format("Table '%s' has %s out of %s partitions unreadable: ", tableName, unreadablePartitionsSkipped, partitionBatch.size()));
            for (Entry<String, Set<String>> entry : partitionsNotReadable.entrySet()) {
                warningMessage.append(String.join(", ", entry.getValue())).append("... are due to ").append(entry.getKey()).append(". ");
            }
            warningCollector.add(new PrestoWarning(PARTITION_NOT_READABLE, warningMessage.toString()));
        }
        return results.build();
    });
    return concat(partitionBatches);
}